jmutai/k8s-pre-bootstrap

Ansible role to setup Kubernetes Requirements before kubeadm init

k8s-pre-bootstrap's Issues

Typo in main.yaml

Hello,

There is typo in tasks/main.yaml Line 32.

- name: Include task to configure containerd container runtime
  include: setup_crio.yml
  when: container_runtime == "containerd"

This should be

-  name: Include task to configure containerd container runtime
   include: setup_containerd.yml
   when: container_runtime == "containerd"

At the beginning of README, describe how to run playbooks as root and sudo user

Take out selinux off and yum update in separate playbook

Docker failed to start.

Ansible fell over when trying to start the docker daemon.

Starting dockerd revealed the issue:-

sudo dockerd
INFO[2020-12-06T13:14:51.712175688Z] Starting up
INFO[2020-12-06T13:14:51.717032757Z] parsed scheme: "unix"                         module=grpc
INFO[2020-12-06T13:14:51.717061129Z] scheme "unix" not registered, fallback to default scheme  module=grpc
INFO[2020-12-06T13:14:51.717083960Z] ccResolverWrapper: sending update to cc: {[{unix:///run/containerd/containerd.sock 0  <nil>}] <nil>}  module=grpc
INFO[2020-12-06T13:14:51.717094701Z] ClientConn switching balancer to "pick_first"  module=grpc
INFO[2020-12-06T13:14:51.719469363Z] parsed scheme: "unix"                         module=grpc
INFO[2020-12-06T13:14:51.719493234Z] scheme "unix" not registered, fallback to default scheme  module=grpc
INFO[2020-12-06T13:14:51.719507424Z] ccResolverWrapper: sending update to cc: {[{unix:///run/containerd/containerd.sock 0  <nil>}] <nil>}  module=grpc
INFO[2020-12-06T13:14:51.719515835Z] ClientConn switching balancer to "pick_first"  module=grpc
failed to start daemon: error initializing graphdriver: overlay2: the backing xfs filesystem is formatted without d_type support, which leads to incorrect behavior. Reformat the filesystem with ftype=1 to enable d_type support. Backing filesystems without d_type support are not supported.

To fix, remove this from templates/daemon.json.j2

  "storage-driver": "overlay2",
  "storage-opts": [
    "overlay2.override_kernel_check=true"
  ],

CentOS Linux release 7.9.2009 (Core)
3.10.0-1160.6.1.el7.x86_64 #1 SMP Tue Nov 17 13:59:11 UTC 2020 x86_64 x86_64 x86_64 GNU/Linux

There is no folder /etc/docker/ during containerd installation

Error when processing setup_containerd.yml: there is no on host folder and file below
#- name: Set cgroup driver as systemd

ansible.builtin.template:

src: daemon.json.j2

dest: /etc/docker/daemon.json

run_once Configure containerd

If uncomment the line below

• name: Configure containerd
  ansible.builtin.shell: containerd config default > /etc/containerd/config.toml
  #run_once: true

the hosts can not join to cluster with error message:
[preflight] Some fatal errors occurred: [ERROR CRI]: container runtime is not running: output: time="2021-12-18T19:30:25+03:00" level=fatal msg="getting status of runtime: rpc error: code = Unimplemented desc = unknown service runtime.v1alpha2.RuntimeService"

To resolve the issue I commented line run_once

kubeadm init can't override cri-o default subnet

When choosing the engine option cri-o, it sticks the pods subnet to 10.85.0.0/16. It is defined in /etc/cni/net.d/10-crio-bridge.conf.
Running kubeadm init ... with the option --pod-network-cidr="A.B.C.D/X" does not override this default subnet.
Suggestion:
Use a template for 10-crio-bridge.conf with a variable for the desired subnet to be used when initializing the k8s cluster later.